Upregulation of CPT1A Is Essential for the Tumor-Promoting Effect of Adipocytes in Colon Cancer

Home , Etomoxir

University of Kentucky UKnowledge

Markey Cancer Center Faculty Publications Markey Cancer Center

9-10-2020

Xiaopeng Xiong University of Kentucky, [email protected]

Yang-An Wen University of Kentucky, [email protected]

Rachelle Fairchild University of Kentucky, [email protected]

Yekaterina Y. Zaytseva University of Kentucky, [email protected]

Heidi L. Weiss University of Kentucky, [email protected]

See next page for additional authors Follow this and additional works at: https://uknowledge.uky.edu/markey_facpub

Part of the Cancer Biology Commons, Medical Molecular Biology Commons, Medical Toxicology Commons, Oncology Commons, and the Surgery Commons Right click to open a feedback form in a new tab to let us know how this document benefits ou.y

Repository Citation Xiong, Xiaopeng; Wen, Yang-An; Fairchild, Rachelle; Zaytseva, Yekaterina Y.; Weiss, Heidi L.; Evers, B. Mark; and Gao, Tianyan, "Upregulation of CPT1A Is Essential for the Tumor-Promoting Effect of Adipocytes in Colon Cancer" (2020). Markey Cancer Center Faculty Publications. 153. https://uknowledge.uky.edu/markey_facpub/153

This Article is brought to you for free and open access by the Markey Cancer Center at UKnowledge. It has been accepted for inclusion in Markey Cancer Center Faculty Publications by an authorized administrator of UKnowledge. For more information, please contact [email protected]. Upregulation of CPT1A Is Essential for the Tumor-Promoting Effect of Adipocytes in Colon Cancer

Digital Object Identifier (DOI) https://doi.org/10.1038/s41419-020-02936-6

Notes/Citation Information Published in Cell Death & Disease, v. 11, issue 9, article no: 736.

This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Authors Xiaopeng Xiong, Yang-An Wen, Rachelle Fairchild, Yekaterina Y. Zaytseva, Heidi L. Weiss, B. Mark Evers, and Tianyan Gao

This article is available at UKnowledge: https://uknowledge.uky.edu/markey_facpub/153 Xiong et al. Cell Death and Disease (2020) 11:736 https://doi.org/10.1038/s41419-020-02936-6 Cell Death & Disease

ARTICLE Open Access Upregulation of CPT1A is essential for the tumor- promoting effect of adipocytes in colon cancer Xiaopeng Xiong1,Yang-AnWen1, Rachelle Fairchild1, Yekaterina Y. Zaytseva 2,HeidiL.Weiss1,B.MarkEvers 1,3 and Tianyan Gao 1,4

Abstract Colon tumors grow in an adipose tissue-enriched microenvironment. Locally advanced colon cancers often invade into surrounding adipose tissue with a direct contact with adipocytes. We have previously shown that adipocytes promote tumor growth by modulating cellular metabolism. Here we demonstrate that carnitine palmitoyltransferase I (CPT1A), a key enzyme controlling fatty acid oxidation (FAO), was upregulated in colon cancer cells upon exposure to adipocytes or fatty acids. In addition, CPT1A expression was increased in invasive tumor cells within the adipose tissue compared to tumors without direct contact with adipocytes. Silencing CPT1A abolished the protective effect provided by fatty acids against nutrient deprivation and reduced tumor organoid formation in 3D culture and the expression of genes associated with cancer stem cells downstream of Wnt/β-catenin. Mechanistically, CPT1A-dependent FAO promoted the acetylation and nuclear translocation of β-catenin. Furthermore, knockdown of CPT1A blocked the tumor-promoting effect of adipocytes in vivo and inhibited xenograft tumor initiation. Taken together, our ﬁndings identify CPT1A-depedent FAO as an essential metabolic pathway that enables the interaction between adipocytes and colon cancer cells. 1234567890():,; 1234567890():,; 1234567890():,; 1234567890():,;

Introduction understood how the direct interaction with adipocytes Altered metabolism has been recognized as a common alters tumorigenic properties of cancer cells. hallmark of cancer1,2. Emerging evidence indicates that Recently, it has been shown that lipids produced in the tumor microenvironment (TME) is essential in adipocytes can be transferred to cancer cells to promote shaping the landscape of cancer metabolism3. A profile of tumor growth in ovarian cancer models, suggesting that the components in the TME revealed that adipocytes local adipose tissues may have a direct role in supporting constitute a major cell type that is abundantly associated cancer cells7. Additionally, we reported that the uptake of with tumor cells4. Metastatic colon cancer cells often fatty acids from adipocytes allows colon cancer cells to encounter adipocytes as they first disseminate from their survive nutrient deprivation conditions by upregulating primary tumor site. Previous studies indicated that mitochondrial fatty acid oxidation (FAO)8. This transfer cancer-associated adipocytes promote tumor growth and of fatty acids from adipocytes to cancer cells has been progression by secreting growth factors and proin- demonstrated in breast and melanoma cancer models as flammatory cytokines into the TME5,6. However, it is less well9,10. Interestingly, the presence of cancer cells stimulates the release of fatty acids by promoting lipolysis in adipocytes, thus indicating a two-way communication between cancer cells and adipocytes in the TME7,8. Fur- Correspondence: Tianyan Gao ([email protected]) thermore, treatment with fatty acids enhances the 1Markey Cancer Center, University of Kentucky, Lexington, KY 40536-0679, USA expression of genes associated with colon cancer stem 2 Department of Toxicology and Cancer Biology, University of Kentucky, cells (CSCs) and suppresses genes associated with Lexington, KY 40536-0679, USA 8 Full list of author information is available at the end of the article intestinal epithelial cell differentiation . This finding is Edited by A. Stephanou

© The Author(s) 2020 Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a linktotheCreativeCommons license, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/.

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 2 of 16

consistent with the notion of CSC plasticity in that non- to the normal blood glucose level) supplemented with CSCs are capable of converting to CSCs given the right 10% lipoprotein-deficient bovine serum (Alfa Aesar, – cue presented by the TME11 13. Taken together, these J65182AMG). The octanoate concentration (3 mM) used studies suggest that the close interaction between adipo- for treating cells was based on the amount of octanoate cytes and cancer cells plays an important role in reg- needed to induce the expression genes related to “fatty ulating cancer metabolism. acid metabolic process” as described previously16. Although the transfer of fatty acids from adipocytes to – cancer cells has been confirmed in several studies7 10, Isolation of human mature adipocytes the molecular mechanism underlying fatty acids- Human omental or mesenteric fat tissues were collected dependent metabolic regulation in cancer cells remains from colon cancer patients undergoing surgery at the elusive. Here we further determined the role of CPT1A, a University of Kentucky Markey Cancer Center. The pro- rate-limiting enzyme required for mitochondrial FAO, in cess for patients’ material collection was approved by the mediating the tumor-promoting effect of adipocytes in University of Kentucky’sOffice for the Protection of colon cancer. Using primary colon cancer cells, 3D Human Subjects. The isolation of adipocytes was carried tumor organoids and in vivo xenograft models, we out as described previously8. Equal amount of adipocytes showed that uptake of fatty acids promotes the expres- were used as determined by the packed cell volume in all sion of CPT1A through the activation of PPARδ.Con- experiments (the density of purified adipocytes used is sequently, knockdown of CPT1A attenuated fatty acid ~1–2×106 cells/ml). utilization and eliminated the pro-survival advantage provided by adipocytes. In addition, we identified Immunohistochemical (IHC) staining β-catenin acetylation as a novel mechanism connecting Paraffin-embedded colon cancer patient specimens upregulation of FAO with increased Wnt/β-catenin sig- were obtained from the Biospecimen Procurement and naling. Together, results from our study provided new Translational Pathology Shared Resource Facility of the mechanistic insights into the tumor-promoting effect of Markey Cancer Center. The diagnosis and staging of each adipocytes in colon cancer. cancer case were confirmed by pathologist. IHC staining of paraffin-embedded tissue sections was performed as Materials and methods previously described8,17. The CPT1A antibody (#12252) Cells and reagents was obtained from Cell Signaling. The stained sections Patient-derived colon cancer PT130 cells were estab- were visualized using a Nikon Eclipse 80i upright micro- lished as described previously8,14. Human colon cancer scope. To quantify the relative CPT1A expression levels, SW480 cells were purchased from ATCC. Both cell pixel intensity values were used to define the percentage lines were maintained in DMEM supplemented with 10% of tumor cells with positive staining using the HALO fetal bovine serum (FBS, Sigma-Aldrich) and 1% image analysis platform (Indica Labs). penicillin–streptomycin. The cell lines were authenticated using short tandem repeat (STR) DNA profiling and tested Measurements of cellular acetyl-CoA (Ac-CoA) levels negative for mycoplasma contamination (Genetica). The The cellular levels of Ac-CoA were determined using shRNA-targeting sequences for human CPT1A are as the the Acetyl-Coenzyme A Assay Kit (Sigma-Aldrich). following: 5′- GCCATGAAGCTCTTAGACAAA-3′ (C6), Briefly, cells were cultured in low glucose media supple- and 5′-CGATGTTACGACAGGTGGTTT-3′ (C7); and for mented with 10% FBS. Total of 106 cells were trypsinized, mouse Cpt1a is: 5′-GCTATGGTGTTTCCTACATTA-3′. lysed in perchloric acid and neutralized using KHCO3. The following reagents were obtained from commercial After centrifuging at 13,000 × g for 10 min, supernatants sources as specified below: Acetyl-Coenzyme A Assay Kit, were collected and used for Ac-CoA measurements. The oleic acid (OA) (albumin complex), palmitic acid (PA), protein contents in the pellets were determined using the linoleic acid (LA) (albumin complex), octanoate, BCA Protein Assay Kit. The levels of Ac-CoA were pre- GW501516, GSK3787, and etomoxir (ETO) were from sented as pmol/mg of protein. Sigma-Aldrich; BIODPY 493/503, N-2, and B-27 supple- ment were from Thermo Fisher Scientific. Bovine serum Seahorse extracellular flux analysis albumin (BSA)-conjugated palmitate was prepared The Seahorse XF96 Extracellular Flux Analyzer (Agi- according to the Seahorse protocol (Seahorse Bioscience). lent) was used to measure the respiration activity of colon The concentrations of long chain fatty acids used in the cancer cells as described previously8,14,18. The mito- cell treatment experiments (100–200 μM) are below the chondrial stress tests were performed according to man- average plasma fatty acid concentrations found in healthy ufacturer’s protocol. The relative levels of basal and people15. For fatty acid treatment experiments, cells were maximal respiration were calculated based on OCR data cultured in low glucose media (5 mM, which is close obtained in the Mito stress tests. The FAO assays were

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 3 of 16

performed as previously described with the following were cultured in 3D Matrigel for 3 days and collected for modifications8. Briefly, cells were seeded at the density of RT-PCR14. 3×104 cells per well in a XF96 plate and subsequently incubated in substrate-limited medium (DMEM with Western blot analysis 0.5 mM glucose, 1.0 mM glutamine, 0.5 mM carnitine, Colon cancer cells or tumor organoids were collected and 1% FBS) for ~16 h. Prior to the beginning of FAO and detergent-solubilized cell lysates were obtained as measurements, the medium was switched into FAO assay described previously8,14. The NE-PER Nuclear and Cyto- medium (200 μM palmitate–BSA, 111 mM NaCl, 4.7 mM plasmic Extraction Kit (Thermo) was used to separate KCl, 2.0 mM MgSO4, 1.2 mM Na2HPO4, 2.5 mM glucose, cytoplasmic and nuclear fractions. Equal amounts of total 0.5 mM carnitine, and 5 mM HEPES). FCCP (3 μM), ETO cell lysates were resolved by SDS–PAGE and subjected to (200 μM), and antimycin A (4 μM) were added subse- Western blot analysis. The following antibodies, including quently at the indicated time. For measurements using CPT1A, acetylated-lysine (pan-Lys, #9441) acetyl-histone OA–BSA as substrate, OA–BSA (100 μM) was added to H3 (Lys9) (#9649), acetyl-histone H3 (Lys27) (#8173), the FAO assay medium replacing palmitate–BSA. All histone H3 (#14269), active-β-catenin (#8814), total OCR measurements were normalized to total protein β-catenin (#8480), were purchased from Cell Signaling; contents in each well. The basal, maximum, and reserved the acetyl-α-tubulin (T7451) and β-actin (A1978) anti- FAO were calculated based on OCR measurements upon bodies were from Sigma-Aldrich; and the total α-tubulin the addition of mitochondrial inhibitors. (sc-5286) and Lamin A/C (sc-20681) antibodies were from Santa Cruz. Fatty acid degradation assay Cells were seeded in 96-well plates at a density of 1 × Quantitative RT-PCR 104 cells per well and incubated with OA (200 μM) Total mRNAs were isolated from cells using the RNeasy overnight. The cells were then switched to low glucose Mini Kit (Qiagen). Equal amounts of RNA were subjected medium supplemented with 10% FBS and fixed in to the reverse transcription PCR with the High Capacity paraformaldehyde at indicated time points. To quantify cDNA Reverse Transcription kit (Thermo). The cDNAs the amount of lipids, the fixed cells were stained obtained were subjected to RT-PCR reactions using the with BODIPY 493/503 (1 µg/ml) and the amount of SYBR Green Master Mix (Thermo) and primers listed in BODIPY fluorescence was measured in cell lysates using Supplementary Table S1. To determine the expression of a SpectraMax M5 Microplate Reader. For visualizing human and mouse CPT1A, gene-specific probes were cellular lipid contents, cells were seeded onto coverslips purchased from Thermo and used in RT-PCR reactions in six-well plates, loaded with OA and allowed to unload using Taqman Gene Expression Master Mix (Thermo). for 24–48 h. The cells were then fixedandstainedwith All values were normalized to the level of β-actin. BODIPY 493/503 and DAPI. The immunofluorescence + images were obtained using a Nikon A1 confocal Xenograft tumorigenesis microscope. All animal procedures were done using protocols approved by the University of Kentucky Animal Care and Tumor organoid formation assay Use Committee. Six to eight week-old NOD.Cg-Prkdcscid Tumor organoids generated from Apc/Kras double IL2rgtm1Wjl/SzJ (NSG, The Jackson Laboratory) mice were mutant mice were cultured as described previously8.To used. Both male and female mice of equal numbers were silence Cpt1a expression, tumor organoids were dis- included in each group. Mice were housed in barrier sociated into small cell clusters using TrypLE (Thermo) rooms with 12-h light/dark cycle. For tumor growth assay, and incubated with sh-Cpt1a lentivirus in suspension for control and CPT1A knockdown SW480 cells in 5% 6 h in a 37 °C incubator. Cells were subsequently Matrigel suspension (5 × 105 cells in 100 µl) were mixed embedded in 50% Matrigel in 3D growth medium with 50 µl of freshly isolated human adipocytes (contain (Advanced DMEM/F12 supplemented with 1 × Glutamax, ~100,000 adipocytes) or PBS and inoculated sub- 1 × N-2, 1 × B-27, 1 mM N-acetyl-L-cysteine, and 1% cutaneously. The tumor size was measured every week penicillin/streptomycin), and puromycin was added with a caliper, and the tumor volume was defined as 2 days later to select for stable knockdown cells. To (longest diameter) × (shortest diameter)2/2. At the end of determine the tumor initiation capacity, organoids were experiments, tumors were harvested and subjected to dissociated to single cell suspensions using Accumax mRNA and protein analysis. For tumor initiation assay, (Sigma-Aldrich). Total of 1000 cells per group were 100 or 1000 SW480 cells were mixed with Matrigel and embedded in Matrigel as described above. The number of adipocytes as described above and injected sub- tumor organoids formed after 6 days were counted and cutaneously. The number of tumors formed was deter- analyzed. For gene expression analysis, tumor organoids mined 3 months post injection.

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 4 of 16

Statistical analysis Silencing CPT1A alters cellular metabolism in colon cancer In experiments to assess rate of FAO, relative cell survival, cells mRNA expression, levels of Ac-CoA and colony formation Given that CPT1A is the first rate-limiting enzyme in were summarized using bar graphs and pairwise compar- FAO, we next investigated the effect of silencing CPT1A isons between different conditions were carried out using on cellular metabolism. Stable CPT1A knockdown PT130 two-sample t-tests. For measuring the time course of fatty and SW480 cells were generated using two different len- acid degradation, one-way or two-way analysis of variance tiviral shRNAs (Fig. 2a). To determine the rate of FAO, models with two-way interaction terms for experimental the oxygen consumption rates (OCRs) were measured by factors such as treatment group, cell types and time were using palmitate as the metabolic substrate in control and utilized. A linear mixed model was employed to compare CPT1A knockdown cells with Seahorse XF96 Extra- slope of tumor volume growth curves over time between cellular Flux Analyzer (Fig. 2b–e). We found that the groups. The relative mRNA expression results represent OCRs associated with basal, maximal. and reserved FAO average of three separate RT-PCR experiments with four were significantly decreased in CPT1A knockdown PT130 replicates for each gene in each experiment. All other and SW480 cells (Fig. 2c, e), confirming the role of experiments were repeated three times and results shown CPT1A in controlling mitochondrial FAO. Similar results represent the average of three experiments. Measurements were obtained using OA as the metabolic substrate in of xenograft tumor growth were summarized at each time Seahorse measurements to confirm that knockdown of point of follow-up and analysis was performed using long- CPT1A decreases FAO in colon cancer cells (Supple- itudinal models to account for repeatedly measured tumor mentary Fig. S2). In addition, since FAO and glycolysis are volume over time within each mouse. The stem cell fre- two functionally coupled metabolic processes that can be quencies were calculated by extreme limiting dilution ana- utilized by cancer cells, we determined if knockdown of lysis (ELDA)19 using analytic tools available at http://bioinf. CPT1A alters mitochondria-dependent glucose metabo- wehi.edu.au/ software/elda/. lism using Seahorse analysis (Supplementary Fig. S3). Interestingly, when using glucose as the metabolic sub- Results strate, results from Seahorse Mito Stress Tests showed CTP1A expression is upregulated by fatty acids in colon that the OCRs associated with both basal and maximal cancer mitochondrial respiration as well as ATP production were We have shown previously that fatty acids released by significantly increased in CTP1A knockdown cells (Sup- adipocytes can be taken up by cancer cells to support cell plementary Fig. S3A–D). Moreover, measurements from survival by upregulating FAO8. Here we further investi- Seahorse Glycolysis Tests indicated that the extracellular gated the mechanisms by which fatty acids induce this acidification rate (ECAR) was increased in CTP1A metabolic switch. To this end, we determined the expres- knockdown cells (Supplementary Fig. S3E–H). These sion of CPT1A in tumor specimens obtained from stage IV results are consistent with previous reports that ETO colon cancer patients. Interestingly, the expression of treatment can increase glucose uptake in normal and CPT1A was significantly increased in tumor cells that have malignant cells21,22. Taken together, our results suggest invaded into the omental adipose tissue (adipocyte adja- that downregulation of CPT1A impairs FAO in colon cent) compared to tumor cells without direct contact with cancer cells. However, this decreased FAO can be com- adipocytes (primary tumors) (Fig. 1a–d and Supplementary pensated by increased glucose metabolism. Indeed, we did Fig. S1). To examine the effect of fatty acids on CPT1A not observe significant changes in the rate of cell pro- expression directly, we co-cultured PT130, a patient- liferation in CTP1A knockdown cells when cultured in derived colon cancer cell line, and SW480 cells with adi- regular growth medium containing sufficient supplies of pocytes. The CPT1A expression was markedly increased in glucose (Supplementary Fig. S4). the presence of adipocytes as measured by quantitative RT- PCR in colon cancer cells (Fig. 1e) and Western blotting Knockdown CPT1A impairs fatty acid utilization and cell analysis (Fig. 1f). Additionally, PT130 and SW480 cells survival under nutrient deprivation conditions were treated with OA, PA, or LA, three major fatty acid To address the question if upregulation of CPT1A is species found in human adipose tissues20, and the necessary for cancer cells to utilize the exogenous fatty expression of CPT1A was consistently increased at both acids, control and CPT1A knockdown PT130 and SW480 protein and mRNA levels (Fig. 1f, g). In subsequent cells were treated with OA for 24 h (the loading phase) experiments, OA or PA was used to examine CPT1A- and subsequently cultured in low glucose media for addi- dependent effects in colon cancer cells. Together, these tional 24–48 h (the unloading phase) to allow the pre- results suggest that upregulation of CPT1A may provide a ferential utilization of fatty acids. Upon loading cells with necessary mechanism for cancer cells to adapt a fatty acid- OA, similar levels of lipid droplets accumulation were enriched microenvironment. detected in control and CPT1A knockdown cells using

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 5 of 16

Fig. 1 The expression of CTP1A is upregulated by fatty acids in colon cancer. a–c The expression of CPT1A protein was detected in a stage IV colon cancer patient tissues using IHC staining. The black-boxed and blue-boxed regions shown in a were enlarged and presented in b and c, respectively. Scale bar, 100 μm. d Quantitative analysis of CPT1A expression in specimens from 11 stage IV colon cancer patients. The percentage of invasive tumor cells in areas adjacent to adipocytes with positive CPT1A expression was higher than that of non-invasive primary tumors. Data represents mean ± SD (n = 11, *p < 0.01). e PT130 and SW480 cells were co-cultured with adipocytes isolated from colon cancer patients for 2 days. The levels of CPT1A mRNA expression were analyzed using RT-PCR. Data represents mean ± SD (n = 3, ¶p < 0.0001 and *p < 0.01). f Protein lysates prepared from cells co-cultured with adipocytes as described in e or treated with BSA, OA, PA, and LA (100 μM of each fatty acid species) for 24 h were analyzed for CPT1A protein expression using Western blotting. The relative levels of CPT1A were quantiﬁed by normalizing to β-actin and compared to control or BSA-treated cells. g PT130 and SW480 cells treated with BSA, OA, PA, or LA as described in f were analyzed for the expression of CPT1A mRNA using RT-PCR. Data represents mean ± SD (n = 3, §p < 0.001 and ¶p < 0.0001).

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 6 of 16

Fig. 2 Knockdown of CPT1A inhibits fatty acid oxidation in colon cancer cells. a The expression of CPT1A protein was analyzed in stable control and CPT1A knockdown PT130 and SW480 cells using Western blotting. Two different shRNA targeting sequences (C6 and C7) were used to silence CPT1A in each cell line. β-actin was used as loading controls. b Representative OCR measurements obtained from the FAO tests performed in control (sh-Control) and CPT1A knockdown (sh-CPT1A-C6 and sh-CPT1A-C7) PT130 cells using the Seahorse XF96 Extracellular Flux analyzer. FCCP, ETO, and antimycin A (Anti-A) were added at the indicated points. c Experiments as shown in b were quantiﬁed and the relative levels of OCR associated with basal FAO, FAO capacity, and FAO reserve were calculated based on the measurements obtained from the addition of different compounds. Data represent the mean ± SD (n = 3, ¶p < 0.0001 and *p < 0.01). d Representative OCR measurements obtained from the FAO tests in sh-control, sh- CPT1A-C6, and sh-CPT1A-C7 SW480 cells using the Seahorse XF96 Extracellular Flux analyzer. e The relative levels of OCR associated with basal FAO, FAO capacity, and FAO reserve were quantiﬁed. Data represent the mean ± SD (n = 3, #p < 0.05, *p < 0.01, and §p < 0.001).

BODIPY-493/505 staining (Fig. 3a). However, while for additional 48 h (Fig. 3c–f). Our results showed that BODIPY-staining of lipid droplets diminished in the both co-culturing with adipocytes and OA treatment unloading phase in control cells, significant levels of significantly increased the survival of control PT130 and BODIPY-staining remained in CPT1A knockdown cells SW480 cells under nutrient deprivation conditions. (Fig. 3a). Moreover, the relative lipid content was quanti- However, the pro-survival effect provided by adipocytes fied by measuring BODIPY fluorescence intensity at dif- and fatty acids was largely abolished in CPT1A knock- ferent time points following OA treatment. The time down cells (Fig. 3c–f). Taken together, our results course of fatty acids degradation was significantly delayed demonstrate that uptake of fatty acids renders colon in CPT1A knockdown PT130 and SW480 cells as indicated cancer cells resistant to nutrient deprivation and by increased BODIPY retention compared to control cells this acquired survival advantage relies on CPT1A- (Fig. 3b). mediated FAO. Furthermore, we determined if silencing CPT1A dis- rupts the protective effect provided by fatty acid uptake. Downregulation of Cpt1a reduce cancer stem cell Control and CPT1A knockdown PT130 and SW480 properties in 3D tumor organoids cells were co-cultured with adipocytes or treated with We have previously shown that the presence of adipo- OA for 24 h and subsequently cultured in EBSS buffer cytes increases the expression of genes associated with

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 7 of 16

Fig. 3 (See legend on next page.)

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 8 of 16

(see figure on previous page) Fig. 3 Knockdown of CPT1A impairs fatty acid utilization and cell survival under nutrient deprivation conditions. a Stable control (sh- Control) and CPT1A knockdown (sh-CPT1A-C6 and sh-CPT1A-C7) PT130 and SW480 cells were incubated with OA for 24 h (loading) and subsequently allowed to grow in regular growth medium for additional 24 or 48 h for PT130 and SW480 cells, respectively (unloading). Representative confocal images of cells stained with BODIPY 493/503 (green) and DAPI (blue). Scale bar, 100 μm. b Control and CPT1A knockdown PT130 and SW480 cells were incubated with OA for 24 h (day 1) and subsequently were stained with BODIPY 493/503 at indicated time points (days 2–4). The fluorescence intensity was measured using a fluorescence spectrophotometer as readout for relative lipid contents in cells. Data represent the mean ± SD (n = 3, #p < 0.05, *p < 0.01, and ¶p < 0.0001). c and d Control and CPT1A knockdown PT130 cells were co-cultured with adipocytes c or pretreated with OA d for 24 h and subsequently cultured in EBSS for additional 48 h. The relative cell survival was measured using crystal violet staining. Data represent the mean ± SD (n = 3, #p < 0.05). e and f Control and CPT1A knockdown SW480 cells were co-cultured with adipocytes e or pretreated with OA f for 24 h and subsequently cultured in EBSS for additional 48 h. The relative cell survival was determined. Data represent the mean ± SD (n = 3, #p < 0.05 and §p < 0.001).

cancer stem cells in intestinal tumor organoids derived the expression of Wnt target genes, silencing Cpt1a from Apc/Kras double mutant mice8. Here we investigated expression abolished the effect of fatty acids (Fig. 4g). if upregulation of CPT1A is required to mediate the effect Collectively, our results suggest that fatty acid treatment of adipocytes on promoting cancer stem cell functions. enhances Wnt signaling in tumor organoids using a First, the Apc/Kras tumor organoids were embedded with CPT1A-dependent mechanism. adipocytes in 3D Matrigel or treated with different fatty acids and the expression of Cpt1A was monitored using Fatty acids increase CPT1A expression and Wnt signaling Western blot and RT-PCR analysis. Consistent with through PPARδ in colon cancer results shown in Fig. 1, the presence of adipocytes or fatty To determine the molecular mechanism by which fatty acids increased Cpt1a expression at both protein and acids and adipocytes promote CPT1A expression, we mRNA levels (Fig. 4a, b). Next, we silenced the expression investigated the functional contribution of peroxisome of Cpt1a using lentiviral shRNA in Apc/Kras organoids. proliferator-activated receptor δ (PPARδ). PPARδ is a Interestingly, while control tumor cells formed spherical lipid sensing nuclear receptor that can be activated by organoids in 3D, the Cpt1a knockdown organoids showed long-chain fatty acids to regulate cellular metabolism25. branched phenotype suggesting potential differentiation Previous studies have shown that PPARδ play a critical (Fig. 4c). Moreover, single cell suspensions of control and role in controlling FAO in muscle or adipocytes tissues Cpt1a knockdown cells were seeded in 3D Matrigel and and high-fat diet enhance tumor initiation potential of the numbers of colonies formed were determined after intestinal organoids in by activating PPARδ26,27. To con- 6 days. We found that the ability of Cpt1a knockdown cells firm the involvement of PPARδ in regulating CPT1A, we to form colonies in 3D was significantly decreased whereas showed that treating PT130 cells or Apc/Kras tumor the percentage of organoids with the branched phenotype organoids with PPARδ agonist GW501516 directly sti- was increased (Fig. 4d). Similar reduction in colony for- mulated the expression of CPT1A and PPARGC1A (a mation and increase in differentiation were observed in known PPARδ target gene) (Fig. 5a, b). In addition, fatty Apc/Kras organoids treated with ETO, an inhibitor of acid-induced upregulation of CPT1A was effectively CPT1A (Supplementary Fig. S5A and B). In addition, since blocked by pretreating cells with PPARδ antagonist the 3D growth media are enriched in fatty acids, silencing GSK3787 (Fig. 5c, d). Importantly, inhibition of fatty acid- Cpt1a also reduced the rate of cell proliferation in tumor induced PPARδ activation decreased the expression of organoids (Supplementary Fig. S5C). However, the Wnt-targeting genes (Fig. 5e) in tumor organoids. Col- reduced colony formation was not due to decreased cell lectively, our data indicate that fatty acids stimulate survival as knockdown of Cpt1a had no effect on cell CPT1A expression to enhance FAO and Wnt signaling viability (Supplementary Fig. S5D). mainly through PPARδ activation in colon cancer cells. Consistent with the notion that Wnt signaling activation is required to promote cancer stem cell properties23,24,we Downregulation of CPT1A decreases β-catenin acetylation found that treating tumor organoids with three different and activation fatty acids significantly increased expression of Wnt target We next determined the mechanism underlying fatty genes and decreased genes associated with intestinal epi- acid-induced activation of Wnt signaling. The FAO has thelial cell differentiation (Fig. 4e). Moreover, knockdown been identified as a major carbon source for the pro- of Cpt1a largely reduced the expression of genes associated duction of cellular Ac-CoA, a central metabolite that with CSCs downstream of Wnt/β-catenin and increased regulates protein acetylation16,28.Hereweexamined the expression of differentiation markers (Fig. 4f). Fur- whether CPT1A downregulation alters cellular Ac-CoA thermore, while fatty acid treatment significantly increased levels. As shown in Fig. 6a, knockdown of CPT1A

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 9 of 16

Fig. 4 (See legend on next page.)

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 10 of 16

(see figure on previous page) Fig. 4 CPT1A-dependent fatty acid oxidation promotes Wnt signaling and cancer stem cell properties. a Tumor organoids derived from Apc/ Kras double mutant mice were embedded with adipocytes in 3D Matrigel or treated with BSA, OA, PA, and LA (100 μM each) for 2 days. The levels of Cpt1a expression were analyzed using Western blotting. The relative levels of CPT1A were quantified by normalizing to β-actin and compared to control or BSA-treated cells. b Tumor organoids treated as described in a were analyzed for the expression of Cpt1a mRNA using RT-PCR. Data represents mean ± SD (n = 3, *p < 0.01, §p < 0.001, and ¶p < 0.0001). c The expression of Cpt1a was silenced in Apc/Kras mutant tumor organoids using lentiviral shRNA. Single cell suspensions of control and Cpt1a knockdown cells were seeded in 3D Matrigel. Representative images of control and Cpt1a knockdown tumor organoids are shown after 6 days in culture. Scale bar, 100 μm. d The number of tumor organoids formed and the percentage of organoids showed branching phenotype were quantified (total 1000 cells were seeded per group). Data represent the mean ± SD (n = 3, #p < 0.05). e Tumor organoids treated with BSA, OA, PA, and LA (100 μM each) were analyzed for the expression of Wnt β-catenin target genes (including Lgr5, Axin2, and Tcf7) and genes associated with intestinal epithelial cell differentiation (including Krt20 and Muc2) using RT-PCR. Data represents mean ± SD (n = 3, #p < 0.05, *p < 0.01, §p < 0.001, and ¶p < 0.0001). f Control and Cpt1a knockdown organoids were subseeded and grown in 3D Matrigel for 3 days. The expression of Cpt1a, Wnt/β-catenin target genes (including Lgr5, Axin2, and Myc), and genes associated with intestinal epithelial cell differentiation (including Krt20 and Muc2) was determined using RT-PCR. Data represent the mean ± SD (n = 3, *p < 0.01 and §p < 0.001). g Control and Cpt1a knockdown organoids were treated with OA for 2 days. The mRNA expression of Cpt1a as well as target genes of Wnt/β-catenin (including Lgr5 and Myc) was determined using RT-PCR. Data represent the mean ± SD (n = 3, *p < 0.01). significantly reduced total Ac-CoA levels in both PT130 Furthermore, control and CPT1A knockdown PT130 and SW480 cells. Given that fatty acid-derived Ac-CoA cells treated with PA were subjected to Western blot can be used as substrate for protein acetylation and analysis to determine the functional contribution of acetylation of β-catenin enhances Wnt signaling by CPT1A-dependent FAO in regulating protein acetylation. – promoting β-catenin nuclear localization29 31,we Results showed that PA treatment increased levels of investigated whether the reduced acetyl-CoA produc- active β-catenin as well as acetylated α-tubulin, H3K9, and tion affects β-catenin activity. To this end, endogenous H3K27 in control cells whereas silencing CPT1A blocked β-catenin proteins were immunoprecipitated from the effect of fatty acids on β-catenin activation and protein PT130 cells treated with BSA or OA and the level of acetylation (Fig. 6g). Similar results were obtained in β-catenin acetylation was determined using an control and CPT1A knockdown cells treated with OA acetylated-lysine antibody. Indeed, treatment with OA (Supplementary Fig. S6B). As a control, we treated cells increased β-catenin acetylation (Fig. 6b). Moreover, with octanoate, a medium-chain fatty acid that can enter knockdown of CPT1A markedly reduced levels of mitochondria and undergo FAO independent of β-catenin acetylation in PT130 and SW480 cells sug- CPT1A16. Interestingly, octanoate treatment markedly gesting CPT1A-dependent FAO controls β-catenin increased the levels of acetylated α-tubulin, H3K9, and acetylation (Fig. 6c). Consistent with the notion that H3K27 as well as active β-catenin in both control and acetylation of β-catenin increases its activity, results CPT1A knockdown cells (Fig. 6h). Consistent with a dose- from Western blot analysis showed that the levels of dependent effect of octanoate on modulating protein active β-catenin (non-phosphorylated form) were acetylation16, we showed that treating cells with a lower decreased in CPT1A knockdown PT130 and SW480 concentration of octanoate partially rescued the acetyla- cells (Fig. 6d). As controls, we analyzed the acetylation tion defect observed in CPT1A knockdown cells (Sup- status of α-tubulin, a cytoplasmic substrate, and histone plementary Fig. S6C). Taken together, our results suggest H3, a nuclear substrate, at lysine 9 (H3K9) and lysine 27 the acetyl-CoA derived from long-chain FAO can be used (H3K27) residues. Consistently, the levels of Ac-α- as the substrate for protein acetylation in a CPT1A- tubulin as well as H3K9Ac and H3K27Ac were reduced dependent manner. Because free fatty acids derived from in CPT1A knockdown cells confirming that FAO is adipocytes are in the form of long-chain fatty acids, important for regulating a broad range of protein acet- upregulation of CPT1A allows the cancer cells to produce ylation events (Fig. 6d). Similar decrease in the levels of increasing levels of Ac-CoA via mitochondrial FAO to active β-catenin and acetylated α-tubulin, H3K9 and promote β-catenin activation and Wnt signaling. H3K27 was observed in Cpt1a knockdown tumor organoids (Supplementary Fig. S6A). To further deter- Knockdown of CPT1A inhibits xenograft tumor growth and mine the functional effect of decreased β-catenin acet- tumor initiation in vivo ylation, control and CPT1A knockdown PT130 and To examine the effect of silencing CPT1A on tumor SW480 cells were fractionated into cytoplasmic and growth in vivo, we subcutaneously injected control and nuclear fractions. The amount of β-catenin in the CPT1A knockdown SW480 cells mixed with Matrigel nuclear factions was decreased in CPT1A knockdown alone or in combination with human adipocytes into NSG PT130 and SW480 cells confirming that acetylation of mice and monitor the tumorigenesis process. Consistently β-catenin promotes its nuclear localization (Fig. 6e, f). with the tumor promoting effect of adipocytes, the rate of

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 11 of 16

Fig. 5 Fatty acids activate PPARδ to promote CPT1A expression and Wnt signaling in colon cancer cells. a PT130 cells were treated with PPARδ agonist GW501516 (1 μM) for 24 h. The expression of CPT1A and PPARGC1A was determined using RT-PCR. Data represent the mean ± SD (n = 3, ¶p < 0.0001). b Apc/Kras tumor organoids were treated with PPARδ agonist GW501516 (1 μM) for 48 h. The expression of Cpt1a and Ppargc1a was determined using RT-PCR. Data represent the mean ± SD (n = 3, *p < 0.01 and §p < 0.001). c PT130 cells were treated with OA alone or in combination with PPARδ antagonist GSK3787 (1 μM) for 24 h. The expression of CPT1A was determined using RT-PCR. Data represent the mean ± SD (n = 3, #p < 0.05,§p < 0.001, and ¶p < 0.0001). d PT130 cells were treated with PA alone or in combination with PPARδ antagonist GSK3787 (1 μM) for 24 h. The expression of CPT1A and PPARGC1A was determined using RT-PCR. Data represent the mean ± SD (n = 3, §p < 0.001 and ¶p < 0.0001). e Apc/Kras tumor organoids were treated with OA alone or in combination with PPARδ antagonist GSK3787 (1 μM) for 48 h. The expression of Cpt1a, Ppargc1a, and Wnt/β-catenin target genes (including Lgr5, Myc, and Axin 2) was determined using RT-PCR. Data represent the mean ± SD (n = 3, #p < 0.05, *p < 0.01, §p < 0.001, and ¶p < 0.0001). tumor growth was largely increased when the control entire tumorigenesis process. Nevertheless, results from SW480 cells were co-injected with adipocytes (Fig. 7a). RT-PCR analysis showed that the expression of CPT1A Although knockdown of CPT1A did not signiﬁcantly and Wnt target genes (including LGR5 and MYC) was decrease the tumor growth rate basally compared to the elevated in control SW480 cells that co-injected with control group, CPT1A-loss blocked the tumor promoting adipocytes and knockdown of CPT1A largely abolished effect of adipocytes in vivo (Fig. 7a). Analysis of tumor this effect (Fig. 7d). tissues collected from four groups of mice revealed that Furthermore, we performed tumor initiation experi- knockdown of CPT1A decreased the amount of active ments by injecting control and CPT1A knockdown β-catenin, although we did not observe the adipocyte- SW480 cells mixed with Matrigel alone or in combination induced activation of β-catenin in tumors derived from with adipocytes into NSG mice at 100 and 1000 cells per control cells (Fig. 7b, c). It is likely that the effect of adi- site. The number of tumors formed was determined after pocytes on promoting protein acetylation gradually 3 months. Based on an ELDA, co-injection of adipocytes diminished as adipocytes did not expect to survive the signiﬁcantly increased the stem cell frequency in control

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 12 of 16

Fig. 6 (See legend on next page.)

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 13 of 16

(see ﬁgure on previous page) Fig. 6 Downregulation of CPT1A decreases cellular levels of acetyl-CoA and protein acetylation. a Control and CPT1A knockdown PT130 and SW480 cells were cultured in low glucose media supplemented with 10% FBS. The levels of acetyl-CoA in control and CPT1A knockdown PT130 and SW480 cells were determined using the Acetyl-CoA Assay kit. Data represent mean ± SD (n = 3, #p < 0.05 and *p < 0.01). b PT130 cells were treated with BSA or OA (100 μM) for 24 h. Cell lysates were immunoprecipitated using the anti-β-catenin antibody and analyzed for the levels of β-catenin acetylation using the acetylated-lysine antibody (Ac-Lys). The blot was stripped and reprobed for total β-catenin in the immunoprecipitates. c Cell lysates from control and CPT1A knockdown PT130 or SW480 cells were immunoprecipitated and analyzed for the levels of β-catenin acetylation using the acetylated-lysine antibody (Ac-Lys). The blot was stripped and reprobed for total β-catenin in the immunoprecipitates. d Cell lysates from control and CPT1A knockdown PT130 or SW480 cells were analyzed for the levels of active-β-catenin as well as acetyl-α-tubulin (Ac-α-tubulin), acetyl- histone at K9 and K27 residues (H3K9Ac and H3K27Ac) using Western blotting. Total β-catenin, α-tubulin, histone H3, and β-actin were used as loading controls. e and f Control and CPT1A knockdown PT130 e and SW480 f cells were fractionated into cytoplasmic and nuclear fractions. The amount of β-catenin in each fraction was determined using the anti-β-catenin antibody. α-Tubulin and lamin A/C were used as controls for the cytoplasmic and nuclear fractions, respectively. g Control and CPT1A knockdown PT130 cells were treated with BSA or PA (100 μM) for 24 h and cell lysates were analyzed for the levels of active-β-catenin as well as Ac-α-tubulin, H3K9Ac, and H3K27Ac using Western blotting. h Control and CPT1A knockdown PT130 cells were cultured under control condition or treated with octanoate (3 mM) for 24 h and cell lysates were analyzed for the levels of active-β-catenin as well as Ac-α-tubulin, H3K9Ac, and H3K27Ac using Western blotting. cells (from 1 in 458 basally to 1 in 102 in the presence of The cancer stem cells, also known as tumor-initiating adipocytes) but not in CPT1A knockdown cells (Fig. 7e). cells, have been implicated in tumor initiation, recurrence, – Taken together, our results suggest that CPT1A is and metastasis34 36. Emerging evidence suggests that required for adipocytes to enhance the tumor initiation Wnt/β-catenin signaling is required for the maintenance potential in vivo. of normal and cancer stem cells37. Despite having acti- In summary, results from this study support a model vating mutations in the Wnt pathway, colon cancer cells in which uptake of fatty acids activates PPARδ-depen- with the highest levels of β-catenin signaling display dent transcription of CPT1A and CPT1A-dependent cancer stem cell properties23. Consequently, the expres- FAO. Subsequently, increased production of Ac-CoA sion levels of β-catenin target genes have been used as a and the acetylation of β-catenin promote Wnt signaling readout for the stemness of cancer stem cells38. Recent and cancer stem cell function. Thus, pharmacological studies have begun to establish the role of FAO in reg- inhibition of CPT1A with ETO or knockdown of ulating normal and cancer stem cell function. For exam- CPT1A expression may block fatty acids-induced ple, it has been shown that CPT1A-dependent FAO is tumor promoting effects in colon cancer (Supplemen- required for the maintenance of tumor initiation cells in tary Fig. S6D). hepatocellular carcinoma as well as normal hematopoietic – and neural stem cells39 41. Moreover, deletion of Cpt1a in Discussion mouse intestine epithelium decreases the number and Several large prospective epidemiological studies have function of normal intestinal stem cells42. Our study provided strong evidence supporting the role of obesity in identiﬁes CPT1A upregulation as a key metabolic altera- promoting colon cancer initiation and progression5,32,33. tion that cancer cells adapt to promote β-catenin acet- However, the molecular mechanism underlying how adi- ylation and activation in an adipocyte-enriched TME. pose tissue and adipocytes support tumor growth and Since uptake of fatty acids induces the acetylation of progression remains largely unknown. We have shown proteins other than β-catenin, additional studies are previously that abundant adipocytes are found in direct needed to determine the functional contribution of these contact with colon cancer cells and uptake of fatty acids acetylation events. It is of particular interest to further released by adipocytes promotes cancer cell survive by investigate the epigenetic alterations associated with his- upregulating mitochondrial FAO8. Results from this study tone acetylation changes induced by fatty acids. demonstrate that the presence of adipocytes or fatty acids In addition, it has been shown recently that CPT1 stimulates the expression of CPT1A by activating PPARδ- complex may support cell proliferation independent of its dependent transcription. Silencing CPT1A expression in ability to control FAO43. Although we show that inhibi- colon cancer cells blocks the cell survival advantage tion of FAO is coupled with decreased cell survival and provided by adipocytes as a result of decreased fatty acid β-catenin activation in CPT1A knockdown cells, it is degradation via FAO. In addition, CPT1A downregulation possible that CPT1A mediates the tumor promoting induces differentiation of tumor organoids grown in 3D effects of adipocytes using a FAO-independent mechan- and attenuates the effect of fatty acids on promoting the ism. Moreover, a recent study reported that circulating expression of cancer stem cell-associated genes. Impor- fatty acid-binding protein released by adipose tissue tantly, CPT1A expression is required for adipocytes to (A-FABP or FABP4) promotes breast cancer stemness by promote tumor growth and initiation in vivo. activating the IL-6/STAT3/ALDH1 pathway44. The

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 14 of 16

Fig. 7 (See legend on next page.)

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 15 of 16

(see ﬁgure on previous page) Fig. 7 CPT1A is required to mediate the tumor promoting effect of adipocytes in vivo. a Control and CPT1A knockdown SW480 cells were mixed with Matrigel alone or in combination with adipocytes and injected subcutaneously into NSG mice. The size of the tumors was measured every 5 days for 51 days. Data represent the mean ± SEM (n = 6, for sh-Control and sh-CPT1A group; and n = 7 for sh-Control + adipocytes and sh- CPT1A + adipocytes group, #p < 0.05 and *p < 0.01; NS = not signiﬁcant). b Tumor tissues from three mice of each group were analyzed for the levels of CPT1A, active-β-catenin, and β-catenin using Western blotting. c Quantitative analysis of relative active-β-catenin levels in xenograft tumors from four different groups of mice. The levels of active-β-catenin were normalized to total β-catenin in each sample. Data represent the mean ± SD (n = 3, *p < 0.01). d Tumor tissues from three mice of each group were analyzed for the expression of CPT1A, LGR5, and MYC using RT-PCR. Data represent the mean ± SD (n = 3, #p < 0.05, *p < 0.01, §p < 0.001, and ¶p < 0.0001). e Tumor initiation experiments were performed using control and CPT1A knockdown SW480 cells. Cells were mixed with Matrigel alone or in combination with adipocytes and injected into NSG mice at 100 or 1000 cells per site and total eight injections were used for each cell group. The number of tumors formed was determined 3 months post inoculation. The stem cell frequency was calculated using extreme limiting dilution analysis (ELDA) (&p = 0.01, comparing sh-control group with sh-control + adipocytes group; and **p = 0.0003, comparing sh-control + adipocytes group with sh-CPT1A + adipocytes group).

importance of FABP4 in facilitating the transport of fatty References acids from adipocytes to cancer cells has also been shown 1. Cairns,R.A.,Harris,I.,McCracken, S. & Mak, T. W. Cancer cell metabolism. Cold 7,8 Spring Harb. Symp. Quant. Biol. 76, 299–311 (2011). in ovarian and colon cancers . Results from our study 2. Ward,P.S.&Thompson,C.B.Metabolicreprogramming:acancerhallmark here provide strong evidence supporting a role of CPT1A even Warburg did not anticipate. Cancer Cell 21, 297–308 (2012). in controlling the production of signaling metabolites to 3. Anastasiou,D.Tumourmicroenvironment factors shaping the cancer metabolism landscape. Br. J. Cancer 116,277–286 (2017). mediate the communication between adipocytes and 4. Pearce,O.M.T.etal.Deconstructionof a metastatic tumor microenvironment colon cancer cells. Future studies are needed to further reveals a common matrix response in human cancers. Cancer Discov. 8, determine the role of FABP4 in regulating cancer stem- 304–319 (2018). 5. Nieman,K.M.,Romero,I.L.,VanHouten,B.&Lengyel,E.Adiposetissueand ness in colon cancer. adipocytes support tumorigenesis and metastasis. Biochim. Biophys. Acta In summary, results from our study identify CPT1A as a 1831,1533–1541 (2013). key regulator that connects adipocyte-mediated regulation 6. Park,J.,Morley,T.S.,Kim,M.,Clegg,D.J.&Scherer,P.E.Obesityandcancer- mechanisms underlying tumour progression and recurrence. Nat Rev Endo- of cellular metabolism to Wnt signaling in colon cancer crinol 10,455–465 (2014). cells. Given the preferential upregulation of CPT1A by 7. Nieman,K.M.etal.Adipocytespromoteovariancancermetastasisandpro- adipocytes, our findings indicate that inhibition of CPT1A vide energy for rapid tumor growth. Nat. Med. 17,1498–1503 (2011). 8. Wen, Y. A. et al. Adipocytes activate mitochondrial fatty acid oxidation and may provide an effective approach to block the tumor autophagy to promote tumor growth in colon cancer. Cell Death Dis. 8, e2593 promoting effects of adipocytes in colon cancer. (2017). 9.Wang,Y.Y.etal.Mammaryadipocytesstimulatebreastcancerinvasion Acknowledgements through metabolic remodeling of tumor cells. JCI Insight 2, e87489 (2017). This work was supported by R01CA133429 (T.G.), R01CA208343 (B.M.E. and T.G.) 10. Zhang, M. et al. Adipocyte-derived lipids mediate melanoma progression via and a pilot grant from P20GM121327 (University of Kentucky Center for Cancer FATP proteins. Cancer Discov. 8,1006–1025 (2018). and Metabolism). The studies were conducted with support provided by the 11. Batlle, E. & Clevers, H. Cancer stem cells revisited. Nat. Med. 23,1124–1134 Redox Metabolism, Biospecimen Procurement and Translational Pathology, (2017). Cancer Research Informatics, Flow Cytometry and Immune Monitoring, and 12. Shimokawa, M. et al. Visualization and targeting of LGR5(+)humancolon Biostatistics and Bioinformatics Shared Resource Facilities of the University of cancer stem cells. Nature 545,187–192 (2017). Kentucky Markey Cancer Center (P30CA177558) at the University of Kentucky. 13. de Sousa e Melo, F. et al. A distinct role for Lgr5(+) stem cells in primary and metastatic colon cancer. Nature 543, 676–680 (2017). Author details 14. Wen, Y. A. et al. Downregulation of SREBP inhibits tumor growth and 1Markey Cancer Center, University of Kentucky, Lexington, KY 40536-0679, USA. initiation by altering cellular metabolism in colon cancer. Cell Death Dis. 2Department of Toxicology and Cancer Biology, University of Kentucky, 9, 265 (2018). fi Lexington, KY 40536-0679, USA. 3Department of Surgery, University of 15. Abdelmagid, S. A. et al. Comprehensive pro ling of plasma fatty acid con- Kentucky, Lexington, KY 40536-0679, USA. 4Department of Molecular and centrations in young healthy Canadian adults. PLoS ONE 10, e0116195 (2015). Cellular Biochemistry, University of Kentucky, Lexington, KY 40536-0679, USA 16. McDonnell, E. et al. Lipids reprogram metabolism to become a major carbon source for histone acetylation. Cell Rep. 17, 1463–1472 (2016). 17. Liu, J. et al. Loss of PHLPP expression in colon cancer: role in proliferation and Conflict of interest tumorigenesis. Oncogene 28,994–1004 (2009). The authors declare that they have no conflict of interest. 18. Xiong, X. et al. PHLPP regulates hexokinase 2-dependent glucose metabolism in colon cancer cells. Cell Death Discov. 3, 16103 (2017). Publisher’s note 19. Hu,Y.&Smyth,G.K.ELDA:extremelimiting dilution analysis for comparing Springer Nature remains neutral with regard to jurisdictional claims in depleted and enriched populations in stem cell and other assays. J. Immunol. published maps and institutional affiliations. Methods 347,70–78 (2009). 20. Iggman, D., Arnlov, J., Cederholm, T. & Riserus, U. Association of adipose tissue Supplementary Information accompanies this paper at (https://doi.org/ fatty acids with cardiovascular and all-cause mortality in elderly men. JAMA 10.1038/s41419-020-02936-6). Cardiol. 1,745–753 (2016). 21. Lionetti,V.,Stanley,W.C.&Recchia,F. A. Modulating fatty acid oxidation in Received: 17 January 2020 Revised: 21 August 2020 Accepted: 27 August heart failure. Cardiovasc. Res. 90,202–209 (2011). 2020 22. Schlaepfer, I. R. et al. Lipid catabolism via CPT1 as a therapeutic target for prostate cancer. Mol. Cancer Ther. 13, 2361–2371 (2014).

Ofﬁcial journal of the Cell Death Differentiation Association Xiong et al. Cell Death and Disease (2020) 11:736 Page 16 of 16

23. Vermeulen, L. et al. Defining stem cell dynamics in models of intestinal tumor 34. Kreso, A. et al. Self-renewal as a therapeutic target in human colorectal cancer. initiation. Science 342,995–998 (2013). Nat. Med. 20,29–36 (2014). 24. Vermeulen, L. et al. Wnt activity defines colon cancer stem cells and is 35. Kreso, A. et al. Variable clonal repopulation dynamics influence chemotherapy regulated by the microenvironment. Nat. Cell Biol. 12,468–476 (2010). response in colorectal cancer. Science 339,543–548 (2013). 25. Reilly, S. M. & Lee, C. H. PPAR delta as a therapeutic target in metabolic disease. 36. Merlos-Suarez, A. et al. The intestinal stem cell signature identifies col- FEBS Lett. 582,26–31 (2008). orectal cancer stem cells and predicts disease relapse. Cell Stem Cell 8, 26. Furnsinn, C., Willson, T. M. & Brunmair, B. Peroxisome proliferator-activated 511–524 (2011). receptor-delta, a regulator of oxidative capacity, fuel switching and cholesterol 37. Clevers, H. Wnt/beta-catenin signaling in development and disease. Cell 127, transport. Diabetologia 50,8–17 (2007). 469–480 (2006). 27. Beyaz, S. et al. High-fat diet enhances stemness and tumorigenicity of 38. Zeuner, A., Todaro, M., Stassi, G. & De Maria, R. Colorectal cancer stem cells: intestinal progenitors. Nature 531,53–58 (2016). from the crypt to the clinic. Cell Stem Cell 15,692–705 (2014). 28. Pietrocola, F., Galluzzi, L., Bravo-San Pedro, J. M., Madeo, F. & Kroemer, G. Acetyl 39. Chen, C. L. et al. NANOG metabolically reprograms tumor-initiating stem-like coenzyme A: a central metabolite and second messenger. Cell Metab. 21, cells through tumorigenic changes in oxidative phosphorylation and fatty acid 805–821 (2015). metabolism. Cell Metab. 23, 206–219 (2016). 29. Chocarro-Calvo, A., Garcia-Martinez, J. M., Ardila-Gonzalez, S., De la Vieja, A. & 40. Ito, K. et al. A PML-PPAR-delta pathway for fatty acid oxidation reg- Garcia-Jimenez, C. Glucose-induced beta-catenin acetylation enhances Wnt ulates hematopoietic stem cell maintenance. Nat. Med. 18, 1350–1358 signaling in cancer. Mol. Cell 49,474–486 (2013). (2012). 30. Levy, L. et al. Acetylation of beta-catenin by p300 regulates beta-catenin-Tcf4 41. Knobloch, M. et al. A fatty acid oxidation-dependent metabolic shift regulates interaction. Mol. Cell Biol. 24, 3404–3414 (2004). adult neural stem cell activity. Cell Rep. 20,2144–2155 (2017). 31. Hoffmeyer, K., Junghans, D., Kanzler, B. & Kemler, R. Trimethylation and acet- 42. Mihaylova, M. M. et al. Fasting activates fatty acid oxidation to enhance ylationofbeta-cateninatlysine49representkeyelementsinESCplur- intestinal stem cell function during homeostasis and aging. Cell Stem Cell 22, ipotency. Cell Rep. 18, 2815–2824 (2017). 769–778 e764 (2018). 32. Yehuda-Shnaidman, E. & Schwartz, B. Mechanisms linking obesity, inflamma- 43. Yao, C. H. et al. Identifying off-target effects of etomoxir reveals that carnitine tion and altered metabolism to colon carcinogenesis. Obes. Rev. 13, palmitoyltransferase I is essential for cancer cell proliferation independent of 1083–1095 (2012). beta-oxidation. PLoS Biol. 16, e2003782 (2018). 33. Calle, E. E., Rodriguez, C., Walker-Thurmond, K. & Thun, M. J. Overweight, 44. Hao, J. et al. Circulating adipose fatty acid binding protein is a new link obesity, and mortality from cancer in a prospectively studied cohort of U.S. underlying obesity-associated breast/mammary tumor development. Cell adults. N.Engl.J.Med.348,1625–1638 (2003). Metab. 28,689–705.e685 (2018).

Ofﬁcial journal of the Cell Death Differentiation Association